Podcast: Deploying generative AI on the plant floor

Dale Tutt is the vice president of industry strategy at Siemens Digital Industry Software. He spent more than 30 years in the aerospace and defense industry designing and building new aircraft and new spaceships before joining Siemens. Aaron Merkin is chief technology officer (CTO) of Fluke Reliability Solutions. In his current position, Aaron is responsible for developing and executing IIoT strategy and leading the technology team in the continued creation of innovative solutions for customers. Dale and Aaron recently spoke with Smart Industry managing editor Scott Achelpohl about artificial intelligence’s impact on the manufacturing industry.

Below is an excerpt from the podcast:

IW: Aaron and Dale, I want to ask you about generative AI, specifically the large language models like open AI. How might they be deployed for many manufacturing operations directly on the plant floor, for example?

AM: So, we're a reliability-focused organization. If you think about the plant floor, you've got your maintenance technicians and you have your operators and the equipment. And what we see with our focus on reliability is the use of generative AI to address the skills shortages, to some extent labor shortages, but more importantly, the eroding of domain expertise and skills in industrial facilities. And as Dale alluded to, creating copilot-like capabilities where you're able to take a relatively junior engineer or technician and through the kind of adaptive coaching and directions to them, have them perform much more of the complex maintenance tasks than perhaps they would be able to with just their limited amount of experience. So, there's a bit of taking the graying workforce, somebody who's been working with machines for 20 or 30 years, and trying to codify that knowledge, either through looking backwards at the history of work orders performed on assets, or through more direct captive technologies and how they perform their work, create a baseline definition of how the guided work instructions should be provided to reliability engineers, but then make those instructions themselves.

I think the key part that is unique to generative AI capabilities that are coming out now is that ability to adapt in real time the type of instructions that you give. So, you may have an individual who, for example, constantly forgets that the bolt is left-hand threaded versus right-hand threaded. So, for that individual, your work instructions and the coaching you're going to get is to remind them before they go strip the bolt that this is a left-handed one. Whereas, perhaps, you have somebody who's been there 10 years and they still perform the same maintenance routine, but they don't need that extra noise and distraction of being reminded of things. I think the ability to codify deep domain expertise that has been in facilities, simplified and translated for individuals and personalized to make individual people much more productive is where we're seeing opportunity in the reliability space.

DT: I'm very excited about the direction that we're seeing generative AI and where AI can really take us in our manufacturing environments. And the example that Aaron gave relative to how you optimize, say, work instructions depending on the worker’s technical capabilities, maybe how often they've done that, I think that's a great area. One of the things I really want to talk about here is how Siemens’ work with AWS has really been democratizing access to AI and generative AI so that we can deploy it out to more people more quickly. And by bringing our expertise together, we're really helping customers scale up the implementation of some of this technology and transforming their businesses. I think as we see this being deployed in the creation of work instructions, helping out with optimizing quality control processes, maybe even optimizing your plan that you laid out, we're trying to bring this capability to companies faster and on top of their existing solutions.

At CES, at the Consumer Electronics Show a few months ago, we announced this partnership with AWS to integrate the generative AI solutions from Amazon Bedrock into our Mendix low-code platform so that people can go in and bring some of these applications to the marketplace faster and to companies faster, building on their existing solutions, and they're extending it. So, the ability to have that, I think it's a real opportunity to really make it more accessible to a lot of our customers.

About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.